Insurers have big ambitions for artificial intelligence (AI), which we define as computer systems that can sense their environment, then think, learn and take action in response. Here are the top findings from insurers who participated in our survey of US companies actively using AI: the top benefits they see, the top obstacles they report and how to overcome those obstacles.
Among insurers who have already invested in AI, many are reaping big benefits: Almost two-thirds report success in using AI to create a better customer experience (CX). Nearly half say AI is helping improve decision-making. As reported by PwC specialists working with insurers on AI initiatives, companies are increasingly using AI to:
Yet many insurers who seek to deploy AI are getting stuck — usually in the same places.
Insurers are concerned about some of the risks that AI solutions, if not carefully governed, could create: Potential new cybersecurity and privacy threats top the list of AI worries, cited by 42% and 36% of survey respondents, respectively. That may be why PwC’s AI experts report that insurers’ risk and regulatory teams frequently hit the brakes on initiatives.
The problem is not that AI is inherently risky. Instead, the challenge is that even if insurers’ risk and regulatory teams are typically highly sophisticated, they often lack the specialized technology skills and processes required to understand and model AI’s potential impacts.
The lack of AI skills is, of course, a more general challenge, which doesn’t just hold back initiatives for technical reasons. It can create cultural barriers too: Executives may hesitate to greenlight or use tools that they don’t feel comfortable with.
Another common problem is silos, among business lines and between AI and analytics groups. They all need to work together, both to give AI the data it needs and to help initiatives scale up faster.
The following four guidelines can help insurers overcome the obstacles and achieve faster ROI with AI:
Focus on data. Collecting the right data, cleaning it up and standardizing it, and making it available are critical to rapidly and reliably deploying AI.
Centralize capabilities. Bring AI, analytics and automation together to help allocate resources, standardize and utilize data, improve governance and scale solutions.
Think long term. When you start now on developing key capabilities, such as AI upskilling, you’ll likely see benefits for years to come.
Make AI responsible AI. To reduce AI’s risks and make it explainable, apply the responsible AI toolkit.